analytic combinatorics of the mabinogion and ok corral
play

Analytic Combinatorics of the Mabinogion and OK-Corral Urns - PowerPoint PPT Presentation

Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Analytic Combinatorics of the Mabinogion and OK-Corral Urns Philippe Flajolet Based on joint work with Thierry Huillet and Vincent Puyhaubert ALEA08 , Maresias, Brazil, April 2008 1


  1. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Analytic Combinatorics of the Mabinogion and OK-Corral Urns Philippe Flajolet Based on joint work with Thierry Huillet and Vincent Puyhaubert ALEA’08 , Maresias, Brazil, April 2008 1 / 20

  2. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Mabinogion MABINOGION Population of N sheep that bleat either A [ aah ] or B [ eeh ]. At times t = 0 , 1 , 2 , . . . , a randomly chosen sheep bleats and convinces one sheep of the other kind to change its opinion. Time to reach unanimity? Probability that one minority group wins? E.g.: French election campaign (2007): N = 60 , 000 , 000. Probability of reversing of majority of 51%? In the “fair” case ( N / 2 , N / 2), time to reach unanimity? 2 / 20

  3. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral OK Corral OK CORRAL Population of N gangsters of gang either A or B . At times t = 0 , 1 , 2 , . . . , a randomly chosen gangster kills a member of the other group. Time to win? Probability that one minority group survives? E.g.: The OK Corral fight at Tombstone. (Wyatt Earp and Doc Holliday) 3 / 20

  4. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Urn models (1) An urn contains balls of 2 possible colours A fixed set of rules governs the urn evolution: Place � �� � Red (A) Blue (B)  Red (A) Draw :  α β  Blue (B) γ δ Balanced urns: α + β = γ + δ =: σ Classically: β, γ ≥ 0 and σ > 0. Convention: The ball “drawn” is not withdrawn (not taken out)! 4 / 20

  5. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Urn models (2) All (classical) balanced 2 × 2 models are “integrable”! � � α β = An urn ; γ δ = A partial differential operator D = x α +1 y β ∂ x + x γ y δ +1 ∂ y ; = An ordinary nonlinear system { ˙ ˙ X = X α +1 Y β , Y = X γ Y δ +1 } . Refs : [Fl-Ga-Pe’05]; [Fl-Dumas-Puyhaubert’06] [Conrad-Fl’06] [Hwang-Kuba-Panholzer’07+]; [Mahmoud ⋆ ]. Cf also: Janson ⋆ ♥♥♥ 5 / 20

  6. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral § 1. Ehrenfest & Mabinogion 6 / 20

  7. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Ehrenfest � � − 1 1 Ehrenfest’s two chambers E = ; 1 − 1 { ˙ ˙ Formally: D = x ∂ y + y ∂ x ; X = Y ; Y = Y } ; Combinatorics of set partitions: histories from ( N , 0) to ( k , N − k ) are partitions with N − k even classes and k odd classes: � N � P [( N , 0) → ( k , N − k ) , N steps] = n ! · [ z n ] sinh k ( z ) cosh N − k ( z ) N n · k 0 1 2 3 Also: path in a special graph k / N ( N − k ) / N Also: special walks on the interval k − 1 ← − k − → k + 1 7 / 20

  8. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Mabinogion (1) k / N ( N − k ) / N Ehrenfest: k − 1 ← − k − → k + 1 ( N − k ) / N k / N Mabinogion: k − 1 ← − k − → k + 1 + absorption Time-reversal relates M [ N ] and E [ N + 2], with fudge factors Theorem M1. Absorption time T of the Mabinogion urn: � N − 2 � P ( T = n + 1) = N − 1 n ! [ z n ](sinh z ) k − 1 (cosh z ) N − k − 1 . N n +1 k − 1 8 / 20

  9. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Trajectories 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0 0.5 1.0 1.5 2.0 2.5 9 / 20

  10. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Mabinogion (2) Theorem M2. Probability Ω N , k of majority reversal: k = xN is initial # of A’s, with x > 1 2 ; A’s become extinct 1 − lim N log Ω N , k = log 2 + x log x + (1 − x ) log(1 − x ) . N →∞ Proof. Laplace transform + Laplace method (peak at end-point). � ∞ e − z (sinh z ) k − 1 (cosh z ) N − k − 1 dz Ω N , k = · · · 0 � 1 (1 − y ) k − 1 (1 + y ) N − k − 1 dy = · · · � N − 2 � 0 1 2 − N +1 ∼ 2 x − 1 . k − 1 N = 60 , 000 , 000: 51% → 10 − 5 , 215 ; 50.1% → 10 − 54 . 10 / 20

  11. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Mabinogion (3) Theorem M3. Time T till absorption , when k = xN is initial number of A’s, with x < 1 2 : � T − N τ � � t 1 e − w 2 / 2 dw . √ → √ P σ N 2 π −∞ τ ( x ) = 1 1 − 2 x ; σ ( x ) 2 = x (1 − x ) 1 (1 − 2 x ) 2 + 1 2 log 2 log(1 − 2 x ). Proof. Laplace transform � ∞ � e uT � e − z � � k − 1 � � N − k − 1 sinh uz cosh uz = · · · dz E N N 0 √ Characteristic functions u = e it / N . Laplace method (with peak inside the interval) and perturbation. 11 / 20

  12. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Experiments Distribution of absorption time UNFAIR URN FAIR URN 0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 K 1 0 1 2 3 4 Normal � 12 / 20

  13. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Mabinogion (4) Theorem M4. FAIR URN N = ν + ν : time � T till absorption T = n ) ∼ 2 n = 1 ν Ce − t e − e − 2 t ; P ( � 2 ν log ν + t ν. �� 2 j − 1 � 2 � ν/ 2 � ( i ) Exact distribution ∝ Geom . ν j =1 ( ii ) Saddle point for [ z n ](sinh z ) N when n ≈ N log N , with suitable perturbations. � ε ℓ − 1 d � T ∞ ≡ ℓ − 1 / 2 , ε ℓ ∈ Exp(1) . ℓ ≥ 1 [Simatos-Robert-Guillemin’08] [Biane-Pitman-Yor*] 13 / 20

  14. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral § 2. Friedman & OK-Corral 14 / 20

  15. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Friedman � � 0 1 Adverse campaign model: F = 1 0 = A counterbalance of influences [Friedman 1949] • System: { ˙ ˙ X = XY , Y = XY } is exactly solvable Distribution is expressed in terms of Eulerian numbers: rises in perms; leaves in increasing Cayley trees; other urn models [Mahmoud*] [FlDuPu06]. � A n , k u k z n 1 − u A ( z , u ) = n ! = 1 − ue z (1 − u ) n , k � n + 1 � � ( − 1) j ( k − j ) n A n , k = j 0 ≤ j ≤ k 15 / 20

  16. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral OK Corral 100 80 60 Two gangs of m and n gunwomen 40 At any time, one shooter shoots 20 Survival probabilities? 0 0 20 40 60 80 100 Time till one gang wins? ( m , n ) = (50 , 50) 100 [Williams & McIlroy 1998] [Kingman 1999] 80 [Kingman & Volkov 2003] 60 Calamity Jane � Morris & Goscinny c 40 20 0 0 20 40 60 80 100 m + n = 100 16 / 20

  17. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral OK Corral (1) � � � � 0 1 0 − 1 Friedman F = ; OK Corral O = ; i.e., O = −F 1 0 − 1 0 s Time reversal: P O ( m , n ց s ) = m + nP F ( s , 0 → m , n ) Theorem O1. Probability of s survivors of type A, from ( m , n ) � k − 1 �� m + n � m � s ! k m + n − s ( − 1) m − k P O ( m , n , s ) = ( m + n )! s − 1 n + k k =1 Involves generalized Eulerian numbers and expansions + identities. Theorem O2. The probability that first group survives is � m + n � m n � � 1 1 ( − 1) m − k k m + n P O S ( m , n ) = A m + n − k = ( m + n )! ( m + n )! n + k k =1 k =1 17 / 20

  18. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral OK Corral (2) m − n Theorem O3. NEARLY FAIR FIGHTS: If √ m + n = θ , then √ � θ � 1 � 3 1 e − t 2 / 2 dt + O √ P O S ( m , n ) = . n 2 π −∞ Theorem O4. UNFAIR FIGHTS: Probability of survival with m = α n and α < 1 is exponentially small . It is related to the Large Deviation rate for Eulerian statistics. Make use of explicit GFs and usual Large Deviation techniques [Quasi-Powers, shifting of the mean] 18 / 20

  19. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Theorem O5. [Kingman] Number of survivors: √ If ( m − n ) ≫ √ m + n then in probability S → m 2 − n 2 ; If ( m − n ) ≪ √ m + n then � λ 2 √ 3 / 8 1 e − t 2 / 2 dt . P ( S ≤ λ n 3 / 4 ) → √ 2 π −∞ Only dominant asymptotics, indirectly from Kingman et al. Get: Theorem O6. MOMENTS � m + n � m � 1 k m + n − 1 [ f ℓ ( k ) Q ( k ) + g ℓ ( k )] , E [ S ℓ ] = ( − 1) m − k ( m + n )! n − k k =1 k + k ( k − 1) where f ℓ , g ℓ are polynomials and Q ( k ) = k + · · · is Ramanujan’s k 2 = birthday paradox function. � Complex asymptotics ` a la Lindel¨ of–Rice. 19 / 20

  20. Urn models Ehrenfest and Mabinogion Friedman and OK-Corral Conclusions? Analytic combinatorics has something to say about balanced urn models, including some nonstandard ones � = probabilistic approaches [Mahmoud–Smythe–Janson]. Analysis of imbalanced models??? Higher dimensions? “Rather surprisingly, relatively sizable classes of nonlinear systems are found to have an extra property, integrability, which changes the picture completely. Integrable systems [...] form an archipelago of solvable models in a sea of unknown, and can be used as stepping stones to investigate properties of ‘nearby’ non-integrable systems.” Eilbeck, Mikhailov, Santini, and Zakharov (2001) 20 / 20

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend